How to Use Qwen3.8 Max (0902) (batch) in Devin
Qwen3.8 Max 0902 suits complex coding projects and tasks that require several rounds of work. Compared with the original Qwen3.8 Max, Alibaba reports stronger engineering-scale coding and coordination across tools.
Prompts to try with Qwen3.8 Max (0902) (batch) in Devin
Copy an example into Devin, or ask in your own words.
Use Qwen3.8 Max 0902 to investigate this repository's failing tests. Reproduce the first failure, trace its cause, implement a focused fix, and report the checks you ran.
Use Qwen3.8 Max 0902 to review the authentication modules. Map session creation and expiration, identify conflicting behavior, and propose a refactor with explicit acceptance criteria.
Use Qwen3.8 Max 0902 to implement pagination in the existing search endpoint. Follow the project's conventions, cover boundary cases, and verify the response remains compatible with existing callers.
Use Qwen3.8 Max 0902 to compare the API specification with the handlers and tests. List discrepancies with file references, then fix the smallest confirmed mismatch.
Use Qwen3.8 Max 0902 to investigate the supplied error trace against this project. Test competing explanations and return the supported cause, the patch, and any remaining uncertainty.
How to set up Qwen3.8 Max (0902) (batch) in Devin
Pick the way that suits you. Each takes about two minutes.
Use Spiral's MCP server
- Add a custom MCP
Open Customize › MCPs. Click + and choose New MCP. Fill in:
- Sign in to Spiral and approve
In Customize › MCPs, open Spiral and choose Connect. Spiral opens: sign in, pick the organization and click Authorize. No API key.
- Check it
On Spiral's configuration page, click Test tools. Devin lists Spiral's tools.
Use the Spiral skill
- Send Devin this message
Your agent reads Spiral's skill and sets itself up.
set up https://spiral.ai/SKILL.md
- Sign in or paste a key
Approve Spiral in the browser, or give the agent a key from the console.
Use the Spiral CLI
- Install and sign in
Installs the Spiral CLI, signs you in and adds its skill to your coding agents.
npx -y spiral-cli@latest init
Call the Spiral API
- Create an API key
In the console, under API keys.
- Run Qwen3.8 Max (0902) (batch)
One request; the result comes back by polling or a webhook.
curl https://api.spiral.ai/v1/runs \ -H "Authorization: Bearer $SPIRAL_API_KEY" \ -H "Content-Type: application/json" \ -H "Idempotency-Key: $(uuidgen)" \ -d '{ "provider": "openrouter", "model": "qwen/qwen3.8-max-0902:batch", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Devin finds Qwen3.8 Max (0902) (batch) on Spiral, shows the price and runs it.
Qwen3.8 Max (0902) (batch) providers and prices
Spiral runs Qwen3.8 Max (0902) (batch) through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.018424 per run at the default inputs. $1.495 per 1M input tokens, $4.485 per 1M output tokens.
Settings you can ask for
Ask Devin for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Seed | a whole number | ||
| Top k | a whole number from 0 | ||
| Top p | 0 to 1 | ||
| Logprobs | on or off | ||
| Max tokens | 1 to 131072 | The most output tokens to generate. On a reasoning model the reasoning tokens count toward it: a cap too low for the effort can end with no answer (finish_reason length). | |
| Temperature | 0 to 2 | ||
| Top logprobs | 0 to 20 | ||
| Presence penalty | -2 to 2 | ||
| Reasoning effort | minimal, low, medium, high, or xhigh | xhigh | How much the model reasons. |
| Frequency penalty | -2 to 2 |
Every API field of Qwen: Qwen3.8 Max (0902) (batch) is on its tool page: open its card above.
What can you do with Qwen3.8 Max (0902) (batch) in Devin?
Turn a project-level request into code changes checked against the repository's behavior. Qwen3.8 Max 0902 suits work that requires tracing dependencies and revising an implementation from test feedback.
Fix bugs across several modules
Trace a reported failure through related code, reproduce it, and revise the fix using test results.
Plan a large refactor
Map dependencies, identify behavior that must stay intact, and divide the changes into reviewable steps.
Implement features from requirements
Turn a detailed specification into code and check the implementation against explicit acceptance criteria.
Investigate failing tests
Compare failures with the implementation, separate root causes from symptoms, and propose targeted corrections.
Compare long specifications
Identify conflicting requirements and produce a decision list with supporting passages.
Prepare research briefs
Organize supplied research into findings, disagreements, and unanswered questions with source references.
Extract structured records
Turn supplied text into JSON records with explicit rules for missing values and supporting evidence.
Review multi-step workflows
Inspect an execution history, identify the failed decision, and recommend a next action based on recorded results.
More on Qwen3.8 Max (0902) (batch) and Devin
Overview
Qwen3.8 Max (0902) (batch) in other agents
Questions about Qwen3.8 Max (0902) (batch) in Devin
Can Devin use Qwen3.8 Max (0902) (batch)?
Yes. Add Spiral to Devin as an MCP server, and Devin can find and run Qwen3.8 Max (0902) (batch) like any built-in tool.
How much does Qwen3.8 Max (0902) (batch) cost in Devin?
Up to $0.018424 per run at the default inputs. $1.495 per 1M input tokens, $4.485 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Devin plan is separate.
Do I need an OpenRouter account to use Qwen3.8 Max (0902) (batch) in Devin?
No. Spiral runs Qwen3.8 Max (0902) (batch) on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up Qwen3.8 Max (0902) (batch) in Devin?
About two minutes: add a custom MCP, then sign in to Spiral and approve, then check it.
Does Devin wait while Qwen3.8 Max (0902) (batch) runs?
Qwen3.8 Max (0902) (batch) runs in the background. Devin starts the run, checks on it and gives you the text, json when it's ready.
How does Qwen3.8 Max 0902 differ from the original Qwen3.8 Max?
Alibaba describes it as an updated snapshot focused on more complex coding projects, longer autonomous development tasks, and better coordination across tools.
How much text can Qwen3.8 Max 0902 handle?
QwenCloud lists a 1 million token context window and a maximum output of 128k tokens. Treat these as separate limits when planning a long analysis or report.
When is thinking useful with Qwen3.8 Max 0902?
Use thinking for work that requires choosing tools, interpreting their results, and deciding what to do next. A useful example is investigating a bug, testing a proposed fix, and revising it from the test results.
Does valid JSON guarantee consistent results?
No. JSON Object mode can return different key names or value types across requests; QwenCloud recommends JSON Schema mode when the structure must stay fixed.
How should I prompt Qwen3.8 Max 0902 for a complex task?
Give it a concrete deliverable, relevant evidence, constraints, and a clear definition of done. For coding, ask it to use test results to revise its work and report what it actually verified.
Start using Qwen3.8 Max (0902) (batch) in Devin
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into Devin:
set up https://spiral.ai/SKILL.md


